Product system / 4 resources

What the API returns, and which route produces it

Four pages cover the parsing API, PDF-to-Markdown, DocIR JSON, and OCR routing. Each one states the output, the limit, and the control that sits beside the parser.

01 / Product

A document parsing API built for the ingest layer

DocParse gives product teams one authenticated API for turning mixed documents into clean Markdown and structured DocIR JSON. It starts with an inline request, then supports jobs, polling, lifecycle events, cancellation, webhooks, scoped keys, quotas, and artifact purge when the workflow moves into production.

Read page
02 / Product

PDF to Markdown that keeps the document useful

DocParse converts PDFs into Markdown for RAG and LLM workflows while also returning a structured representation of pages and blocks. Digital PDFs use a deterministic text-layer path when possible; scans and complex layouts can be routed to OCR or layout-aware processing under explicit workspace controls.

Read page
04 / Product

Route each document to the parser it actually needs

DocParse uses source type, size, embedded-text evidence, and requested fidelity to choose a deterministic, OCR, layout-aware, image-vision, Office, or external OCR path. The decision is service controlled, recorded with the job, and bounded by tenant capability and cost policy.

Read page

Trial workspace

Send one text-based PDF.

Upload one text-based PDF under 20 MB. Call POST /v1/parse?wait=true with mode=both. Read result.markdown, result.document, and result.manifest. The trial covers 20 documents, 200 MB, and one active job.

Parse one text-based PDF 20 documents · 200 MB · deterministic trial · no payment card